Vitamin D and Macular Thickness in the Elderly: An Optical Coherence Tomography Study

Abstract

Purpose.: Vitamin D insufficiency is associated with age-related macular degeneration. Our objective was to determine whether low serum 25-hydroxyvitamin D (25OHD) concentration was associated with macular thickness among older adults with no signs of macular dysfunction.

Methods.: Sixty-two French older community-dwellers with no patent macular dysfunction (mean ± SD, 71.2 ± 5.0 years; 45.2% female) included in the Gait and Alzheimer Interaction Tracking (GAIT) study (ClinicalTrials.gov number, NCT01315717) were separated into two groups according to serum 25OHD level (i.e., insufficient < 50 nmol/L or sufficient ≥ 50 nmol/L). The macular thickness was measured on 1000 μm central macula with optical coherence tomography, and further binarized according to normal values of macular thickness (i.e., 267.74 μm for males, and 255.60 μm for females). Age, sex, number of comorbidities, cognitive disorders, body mass index, mean arterial pressure, visual acuity, intraocular pressure, serum calcium concentration and season of testing were considered as potential confounders.

Results.: The mean serum 25OHD concentration was 61.2 ± 26.3 nmol/L. Patients with vitamin D insufficiency had a reduced macular thickness compared to those without (232.9 ± 40.4 μm vs. 253.3 ± 32.1 μm, P = 0.042). After adjustment for potential confounders, vitamin D insufficiency was associated with a decreased macular thickness (β = −59.4 μm, P = 0.001). Consistently, the participants with vitamin D insufficiency had a 3.7-fold higher risk of having abnormally low macular thickness compared with those with sufficient 25OHD level (P = 0.042).

Conclusions.: Vitamin D insufficiency was associated with reduced macular thickness among older patients with no patent macular dysfunction. This implies that vitamin D insufficiency may be involved in macular thinning, and provides a scientific base for vitamin D replacement trials in age-related macular degeneration.

Introduction

Beyond its classical contribution to bone health, vitamin D is a secosteroid hormone involved in several target tissues expressing vitamin D receptors, ^1,2 including the retina. ³ Epidemiological literature has recently reported an association between lower 25-hydroxyvitamin D (25OHD) concentrations and impaired visual acuity, ⁴ as well as an association between vitamin D insufficiency and AMD, ^5–8 a clinical condition arising from progressive macular atrophy during aging. No previous epidemiological studies could determine whether AMD precipitated vitamin D insufficiency (due to its clinical expression with consequent loss of function and decreased sun exposure), or whether vitamin D insufficiency had a role in precipitating AMD. To date, no randomized controlled trial has yet explored the benefits of vitamin D supplementation to treat or prevent visual loss and/or AMD. Thus, to infer causality, and before conducting such a trial, it may be contributory to determine whether vitamin D insufficiency is associated with a reduced macular thickness (MT) among participants free of any known macular pathology, and thus independent of any clinical impact. The purpose of our study was to determine whether serum 25OHD was associated with MT measured with optical coherence tomography (OCT) retinal scanning in older adults with no clinical signs of macular dysfunction. 

Materials and Methods

Participants

We studied 73 community dwellers (mean age 70.9 ± 4.9 years; 42.9% female) followed in the Memory Clinic of Angers University Hospital, France, from November 2009 to June 2011 for a subjective memory complaint, and who were recruited into the Gait and Alzheimer Interaction Tracking (GAIT) study (ClinicalTrials.gov number, NCT01315717). The GAIT study is an observational cross-sectional study designed to examine gait in older community dwellers reporting subjective memory complaint. The sampling and data collection procedures have been described elsewhere in detail. ⁹ In summary, subjective memory complaint was documented using the subjective memory complaints questionnaire, ¹⁰ and the main exclusion criteria were aged younger than 60 years, mini-mental state examination score < 10, ¹¹ inability to walk independently, history of stroke, history of any acute medical illness within the past 3 months, current delirium, severe depression, and inability to understand or answer the study questionnaires. For the present analysis, participants were excluded when their refractive status was not fully determined (including the history of refractive status before cataract surgery, when applicable) or when a diagnosis of retinal or macular pathology was made, including advanced AMD (i.e., Age-Related Eye Disease Study categories 3 and 4), diabetic retinopathy, vitreoretinal junction pathology, or macular detachment. Out of 73 participants in the GAIT study, 11 participants were excluded; three due to a diagnosis of macular pathology (one advanced AMD, one macular hole, one history of retinal detachment) and eight because information on refractive status was not fully available. As a result, 62 participants were finally included in this analysis. 

In addition to a full medical examination and blood tests for vitamin D, calcium and albumin concentrations, all included participants underwent a comprehensive ophthalmic clinical examination, evaluating visual acuity, intraocular pressure, fundoscopy, and retinal fundus color imaging. 

Macular Thickness Measurement With OCT

Central MT was determined automatically and analyzed using a spectral-domain high-definition (HD)-OCT device (Cirrus HD-OCT; Carl Zeiss Meditec, Dublin, CA, USA). The pupil was not dilated. In every OCT map, MT detection was performed automatically without manual operator adjustment. HD-OCT (Carl Zeiss Meditec) images were generated using the macular cube 512 × 128 scans. Each image had 5-μm axial and 10-μm transverse resolutions in tissue and consisted of a 512 × 128 volume cube. The scanning area was 6 × 6 mm. The cube was composed of 128 horizontal examination lines of 512 A-scans each. A measure of MT of 1000 μm central retina with HD-OCT (Carl Zeiss Meditec) was performed by an experienced orthoptist for each eye. The MT was estimated based on the 1-mm central retinal thickness area, as described in the Early Treatment Diabetics Retinopathy Study. The average value of two eyes in the same participant was used in our analysis. Abnormally low MT was defined using normal values provided in previous literature (i.e., MT = 267.74 μm for males, and MT = 255.60 μm for females). ¹²  

Serum Vitamin D Insufficiency

Venous blood was collected from resting participants. Serum 25OHD concentration, an effective indicator of vitamin D status, ¹³ was measured by radioimmunoassay (DiaSorin Corp., Stillwater, MN, USA). Intra- and interassay precisions were 5.2% and 11.3%, respectively. Vitamin D insufficiency was defined for 25OHD concentrations < 50 nmol/L according to the definition of the World Health Organization ¹⁴ and the US Institute of Medicine ¹⁵ (to convert to ng/mL, divide by 2.496). All measurements were performed locally at the University Hospital of Angers, France. 

Covariables

The best corrected visual acuity was measured with Monoyer charts and converted into logMAR for statistical analysis purposes. The intraocular pressure in mm Hg was measured by noncontact tonometry (Tonoref; Nidek Co., Ltd., Aichi, Japan). Average values of two eyes in the same participant were used in our analysis. After fundoscopy, images of the retinal fundus were systematically taken via nonmydriatic fundus photography and reexamined post hoc by an experienced ophthalmologist. Evaluation of comorbidities (i.e., diseases lasting at least 3 months or running a course with minimal change, whatever the etiology) was based on self-report and medical record. All participants in the study had a cognitive assessment at the time of inclusion. Cognitive disorders were defined as either mild cognitive impairment or dementia, and were diagnosed using the consensus Winblad et al. ¹⁶ criteria and the criteria of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, ¹⁷ as appropriate. The body mass index (BMI) was calculated as: [weight (kg)/height ² (m²)]. Weight was measured with a beam balance scale, and height with a height gauge. The supine mean arterial pressure (MAP: the average blood pressure that occurs over the entire course of the blood pressure cycle) was calculated from the systolic (SBP) and diastolic blood pressures (DBP) using the following formula: [MAP = (SBP + 2 × DBP)/3]. ¹⁸ The season of evaluation was recorded as follows: spring from March 21 to June 20, summer from June 21 to September 20, fall from September 21 to December 20, winter from December 21 to March 20. Finally, the serum concentration of calcium was measured using automated standard laboratory methods at the University Hospital of Angers, France. Because of the high prevalence of hypoalbuminemia in older adults, calcium values were corrected according to the formula: [corrected calcium value = Ca + 0.02 (46-albumin)]. 

Age, sex, number of comorbidities, cognitive disorders, BMI, MAP, mean visual acuity, intraocular pressure, serum calcium, and season of testing were considered as potential confounders in our analysis. 

Statistical Analysis

The participants' characteristics were summarized using means and SDs or frequencies and percentages, as appropriate. As the number of observations was higher than 40, comparisons were not affected by the shape of the error distribution and no transform was applied. ¹⁹ First, comparisons between participants separated into two groups based on serum 25OHD (i.e., <50 nmol/L or ≥50 nmol/L) were performed using the χ² test or Student's t-test, as appropriate. Second, univariate and multiple linear regressions (i.e., fully adjusted model and backward model) were used to examine the association between vitamin D insufficiency (independent variable) and the MT (dependent variable), while adjusting for potential confounders. Correlation between MT and serum 25OHD concentration used as a quantitative variable was also performed. Finally, logistic regressions were used to examine the association between having abnormally low MT (dependent variable) and participants' characteristics (independent variables). Values of P < 0.05 were considered significant. All statistics were performed using statistical software (SPSS version 19.0; IBM Corporation, Chicago, IL, USA) and Cochrane Review preparation software (Review Manager version 5.1; The Nordic Cochrane Centre, Copenhagen, Denmark). 

Ethics

Participants participating in the study were included after having given their written informed consent for research. The study was conducted in accordance with the ethical standards set forth in the Declaration of Helsinki and the protocol was approved by the University of Angers Ethical Review Committee (CPP Ouest II: 2009–2012). 

Results

Among 62 older participants included in this analysis (mean age, 71.2 ± 5.0 years; 45.2% female; 100% Caucasian), the mean serum 25OHD concentration was 61.19 ± 26.34 nmol/L. Seventeen participants (27.4%) had vitamin D insufficiency. As indicated in Table 1, the participants with vitamin D insufficiency had lower MT than those with 25OHD ≥ 50 nmol/L (232.88 ± 40.41 μm vs. 253.27 ± 32.09 μm, P = 0.042). There were no significant differences for the other clinical characteristics (Table 1). In particular, the mean logMAR visual acuity was 0.07 ± 0.10, with no difference between those with vitamin D insufficiency and those without. Only 14 patients were pseudophakic, with a similar distribution in the group with vitamin D insufficiency and in the group without (P = 0.913). In all, three participants had a history of high myopia of at least −5.0 diopters (D); including two before cataract surgery, and one at the time of assessment. The mean binocular intraocular pressure was 15.96 ± 2.84 mm Hg, with no between-group difference (P = 0.523). Lastly, nine participants had asymptomatic macular drusen, without advanced AMD (P = 0.667 for between-group comparison). 

Table 1

 

Characteristics and Comparison of the Participants

	Total Cohort (n = 62)	Serum 25-Hydroxyvitamin Concentration, nmol/L		P Value*
		<50 (n = 17)	≥50 (n = 45)
Clinical measures
Age, y, mean ± SD	71.23 ± 4.97	71.88 ± 5.17	70.98 ± 4.92	0.527
Female, n (%)	28 (45.2)	8 (47.1)	20 (44.4)	0.854
Number of comorbidities, mean ± SD†	2.18 ± 1.64	2.82 ± 2.07	1.93 ± 1.39	0.055
Cognitive disorders, n (%)‡	35 (56.5)	12 (70.6)	23 (51.1)	0.168
BMI, kg/m2, mean ± SD	25.76 ± 3.72	26.80 ± 4.69	25.36 ± 3.26	0.176
MAP, mm Hg, mean ± SD	97.07 ± 11.98	94.91 ± 14.88	97.90 ± 10.77	0.417
Ophthalmic examination
Visual acuity, logMAR, mean ± SD§	0.07 ± 0.10	0.08 ± 0.11	0.06 ± 0.10	0.630
History of high myopia, n (%)||	3 (4.8)	0 (0.0)	3 (6.7)	0.275
MT, μm,§ mean ± SD	247.68 ± 35.43	232.88 ± 40.41	253.27 ± 32.09	0.042
Intraocular pressure, mm Hg, mean ± SD§	15.96 ± 2.84	15.50 ± 2.11	16.12 ± 3.08	0.523
Drusen detection, n (%)	9 (14.5)	3 (17.6)	6 (13.3)	0.667
Pseudophaky, n (%)	14 (22.6)	4 (23.5)	10 (22.2)	0.913
Serum measures
25OHD concentration, nmol/L, mean ± SD	61.19 ± 26.34	29.00 ± 9.53	73.36 ± 19.42	<0.001
Calcium, mmol/L, mean ± SD	2.37 ± 0.10	2.33 ± 0.08	2.39 ± 0.10	0.032
Season of blood collection				0.221
Spring, n (%)	17 (27.4)	2 (11.8)	15 (33.3)
Summer, n (%)	10 (16.1)	2 (11.8)	8 (17.8)
Autumn, n (%)	31 (50.0)	11 (64.7)	20 (44.4)
Winter, n (%)	4 (6.5)	2 (11.8)	2 (4.4)

 Study participants (n = 62) were separated into two groups based on serum 25-hydroxyvitamin D concentration.

*  Comparisons of participants with normal vitamin D concentrations (i.e., ≥50 nmol/L) with participants with vitamin D insufficiency (i.e., <50 nmol/L) based on χ² test or Mann-Whitney U test, as appropriate.

†  Diseases lasting at least three months or running a course with minimal changes.

‡  Mild cognitive impairment or dementia.

§  Average value of two eyes in the same participant used.

|  | Spherical equivalent refraction of at least −5.0 D; significant P values (i.e., P < 0.05) are indicated in bold.

Figure 1 shows representative examples of central MT obtained from OCT retinal scanning in a participant with normal (A) and insufficient (B) vitamin D status, respectively (P < 0.05). 

Figure 1

 

Representative examples of macular thickness measured with OCT in a participant with sufficient (A) and insufficient (B) vitamin D status. To facilitate the comparison, a box plot of each group point is shown (C). *Macular thickness in the group with vitamin D insufficiency significantly thinner than that in the group with vitamin D sufficiency (P < 0.05).

As illustrated in Table 2, univariate linear regression showed a significant association between vitamin D insufficiency and MT. This association remained significant even after adjustment for all potential confounders (β = −51.74, P = 0.014), and was retained in the backward model (Table 2). Using serum 25OHD concentration as a quantitative variable, we found no significant correlation with MT (r = 0.11, P = 0.410). 

Table 2

 

Univariate and Multiple Linear Regressions

	Central Macular Thickness*
	Unadjusted Model			Fully Adjusted Model			Backward Model
	β	95% CI	P Value	β	95% CI	P Value	β	95% CI	P Value
Vitamin D insufficiency†	–20.38	−40.03;–0.74	0.042	−51.74	−91.61;–11.86	0.014	−59.44	−90.46;–28.42	0.001
Age	−0.91	−2.73;0.92	0.326	0.63	−2.44;3.71	0.667	—	—	—
Female sex	−21.69	−39.04;–4.33	0.015	−23.16	−54.19;7.88	0.133	−24.35	−49.47;0.77	0.057
Number of comorbidities§	−4.91	−10.36;0.54	0.076	−1.64	−10.06;6.78	0.685	—	—	—
Cognitive disorders||	−11.99	−30.03;6.05	0.189	−8.65	−46.32;29.02	0.632	—	—	—
BMI	−2.15	−4.54;0.25	0.078	−0.69	−5.99;4.60	0.784	—	—	—
MAP	−0.02	−0.89;0.84	0.960	1.12	−0.67;2.91	0.203	1.23	−0.10;2.57	0.068
Visual acuity*	21.32	−90.59;133.22	0.703	52.57	−131.28;236.43	0.551	—	—	—
Intraocular pressure*	−3.12	−7.20;0.96	0.131	−5.58	−11.82;0.65	0.073	−4.66	−9.71;0.39	0.068
Serum calcium concentration	42.51	−48.59;133.60	0.354	55.42	−152.63;263.47	0.579	—	—	—

 The table shows the cross-sectional association between macular thickness* (dependent variable) and vitamin D insufficiency† (independent variable), adjusted for potential confounders‡ (n = 62). β, coefficient of regression corresponding to a change in macular thickness.

*  Average binocular measure.

†  Serum 25-hydroxyvitamin D < 50 nmol/L.

‡  Including the influence of seasons.

§  Diseases lasting at least three months or running a course with minimal changes.

|  | Mild cognitive impairment or dementia; significant β values (i.e., P < 0.05) indicated in bold.

Lastly, the logistic regression model showed that the participants with vitamin D insufficiency had a risk multiplied by 3.7 to have abnormally low MT compared to those with sufficient level of 25OHD (P = 0.042; Fig. 2). 

Figure 2

 

Odds ratio (95% confidence interval [CI]) of having abnormally low macular thickness (i.e., MT < 267.74 μm for males, and MT < 255.60 μm for females) according to participants' characteristics (n = 62).

Discussion

The main finding of this OCT study is that, irrespective of all measured potential confounders, vitamin D insufficiency was associated with a thinner central macular thickness among older adults with no patent macular dysfunction. 

To the best of our knowledge, this study is the first to assess and report such an association. This novel finding is consistent with the result of a recent study highlighting an association between vitamin D insufficiency and impaired visual acuity in older adults. ⁴ The authors found among 311 older adults (mean age, 71.7 ± 5.5 years; 39.9% female) that lower serum 25OHD concentrations were associated with reduced vision (P = 0.001). ⁴ Beyond the possible onset of optic neuropathy in the case of low 25OHD status, ²⁰ vitamin D insufficiency-related impaired vision has tentatively been explained by AMD. ^5–8 The firsts to report this association were Parekh and colleagues ⁵ who showed among 7752 adults (mean age, 56.6 years; 56.6% female; 11% with AMD) that the OR for early AMD was 0.64 for participants in the highest versus lowest quintile of serum 25OHD (P trend < 0.001). ⁵ In the second study by Millen and colleagues, ⁶ increased serum 25OHD concentrations were associated with decreased odds of early AMD among 968 women aged <75 years (OR for highest quintile versus lowest quintile = 0.52; P trend = 0.02). However, this result was not confirmed in a population of women aged 75 and older. ⁶ Recently, Seddon and colleagues ²¹ also reported that higher dietary intakes of vitamin D were found in the twins with less severe AMD compared with monozygotic cotwins with more severe AMD (P = 0.01). Although dietary intakes of vitamin D are only an approximate measure of the actual serum vitamin D status, ²² this study suggested a protective effect of vitamin D against the development of AMD. Finally, a case control study comparing 31 patients with AMD and 34 controls reported an association between vitamin D insufficiency <50 nmol/L and late stages of AMD (OR = 3.10, P = 0.031). ⁸ However, because of the cross-sectional design of studies showing an association between vitamin D insufficiency and AMD, and because of two inconclusive studies, ^23,24 it remains thus far impossible to determine whether vitamin D insufficiency had a role in precipitating AMD or whether AMD precipitated vitamin D insufficiency. Importantly, our study, despite its cross-sectional design, highlights an association between vitamin D insufficiency and subclinical macular changes, and thus reinforces the hypothesis of an adverse impact of vitamin D insufficiency on the retina. Consistent is the finding in aged mice that vitamin D administration for 6 weeks significantly reduced the aging processes. ²⁵ Treated mice showed significant reductions in retinal inflammation and levels of amyloid-beta accumulation, together with an improvement of the visual function. This implies that vitamin D insufficiency may be involved in MT thinning, in particular in AMD. Other possible mechanisms have been proposed, including the anti-inflammatory properties of vitamin D. Indeed, several studies have shown epidemiological associations between vitamin D insufficiency and a number of inflammatory diseases including multiple sclerosis or rheumatoid arthritis. ¹ Moreover calcitriol experimentally suppresses antiretinal autoimmunity in experimental autoimmune uveitis induced in mice, through inhibitory effects on the Th17 effector response. ²⁶ Finally, vitamin D may protect against wet AMD with its antiangiogenic properties by inhibiting the proliferation of endothelial cells that express vitamin D receptors. ²⁷ Albert and colleagues ²⁸ have shown, in mice with oxygen-induced ischemic retinopathy and choroidal neovascularization, that a significant reduction in retinal neovascularization was obtained within the calcitriol-treated group compared to control animals. 

The finding that vitamin D insufficiency is associated with reduced MT has interesting potential clinical implications. Even if there was no correlation in our study between MT and serum 25OHD concentration used as a quantitative variable, it is of note that providing a result in terms of linear correlation—in other words, reporting a change in MT related to a change of 1 nmol/L of serum 25OHD—has only poor significance for clinical practice compared with showing an association between vitamin D insufficiency and thinner MT. To the best of our knowledge, there are no clear reference values for a “clinically relevant change” in MT. Of note, the generally accepted and clinically relevant reference value for serum 25OHD concentration is considered to be around 50 nmol/L. ^14,15 In our study, we found a significant decrease of 20.4 μm (8.1%) in MT when comparing vitamin D insufficiency with vitamin D sufficiency (Table 2). Such estimates may help to justify, plan, evaluate, and compare the effectiveness of interventions aiming at preventing macular pathology with vitamin D supplements that would utilize MT change as an outcome measure. 

Some potential limitations of our study should be considered. First, the study cohort was restricted to relatively healthy community-dwelling older participants who might be unrepresentative of the population of all seniors. In particular, only 27.4% of participants had vitamin D insufficiency here, although 40% to 70% of seniors are generally thought to have vitamin D insufficiency in Europe. ² Even if multiple conditions contribute to serum 25OHD status, ¹³ this small prevalence of vitamin D insufficiency was likely explained by the satisfactory nutritional status of participants, as indicated by the mean BMI above 25 kg/m², and by the relatively low morbidity burden (Table 1). Second, the cohort was limited to 62 participants, which may have resulted in lack of statistical power. Third, although we excluded participants with advanced macular pathology that could modify the association, a small proportion of participants with other ocular conditions, such as a history of high myopia or asymptomatic drusen, was still included. Despite these limitations, we were able to show a 3.7-fold higher risk of having abnormally low MT in the case of vitamin D insufficiency among older adults free of clinical retinal diseases. Further well-conducted, multicentric and preferably longitudinal observational cohort studies are needed to corroborate these results on larger samples of participants before recommending vitamin D replacement trials. 

Acknowledgments

The authors thank Melinda Beaudenon, MS, Jennifer Gautier, BS, Romain Simon, MS, and Samuel Thiery, BS, from Angers University Memory Clinic, France; Claire Rabaute and Marielle Chatreaux (orthoptists), Anne Trelohan, MD, Solene Coisy, MD, David Gautier, MD, Stephanie Leruez, MD, Medhi Cherif, MD, and Mathieu Uro, MD, from the Department of Ophthalmology of Angers University Hospital, France, for daily assistance. 

Supported by the French Ministry of Health (Projet Hospitalier de Recherche Clinique National No. 2009-A00533-54). The authors alone are responsible for the content and writing of the paper. 

Disclosure: A. Graffe, None; O. Beauchet, None; B. Fantino, None; D. Milea, None; C. Annweiler, None